Fix VSD cache publication and resolve stub size

[Copilot]

Summary

• Publish virtual stub dispatch cache entries with VolatileStore so initialized entries are safely visible to concurrent readers.
• Correct ResolveStub::size() to return sizeof(ResolveStub) instead of sizeof(LookupStub).

[azure-pipelines]

Azure Pipelines:
Successfully started running 3 pipeline(s).
13 pipeline(s) were filtered out due to trigger conditions.
There may be pipelines that require an authorized user to comment /azp run to run.

[davidwrighton]

@copilot , give the PR a good description.

[dotnet-policy-service]

Tagging subscribers to this area: @agocke
See info in area-owners.md if you want to be subscribed.

[Copilot]

@copilot , give the PR a good description.

Updated the PR description with the two root causes, the fixes, scope separation for the counter work, and validation performed.

[davidwrighton]

@copilot, update the PR description for pull #134226

[azure-pipelines]

Azure Pipelines:
Successfully started running 3 pipeline(s).
13 pipeline(s) were filtered out due to trigger conditions.
There may be pipelines that require an authorized user to comment /azp run to run.

[Copilot]

Copilot review overview

🔵 Needs a closer look

The changes touch low-level VM concurrency and stub code size accounting where subtle memory-ordering or codegen interactions merit final human validation.

Review effort: Lite
Findings: None

What changed in this PR

This PR makes two targeted correctness fixes in CoreCLR’s Virtual Stub Dispatch (VSD) implementation: it strengthens the publication semantics for dispatch cache entries to support concurrent readers, and it corrects the reported size of the AMD64 resolve stub to match the actual stub type.

Changes:

• Publish DispatchCache entries with VolatileStore (paired with existing VolatileLoad reads) to ensure initialized cache entries become safely visible to concurrent lookups.
• Fix ResolveStub::size() on AMD64 to return sizeof(ResolveStub) (matching other architectures and ensuring instruction-cache flush / logging uses the correct byte size).

File	Description
src/​coreclr/​vm/​virtualcallstub.h	Uses VolatileStore(&cache[idx], elem) for cache entry publication to match VolatileLoad readers.
src/​coreclr/​vm/​amd64/​virtualcallstubcpu.hpp	Corrects ResolveStub::size() to return sizeof(ResolveStub) instead of sizeof(LookupStub).

[jkotas]

Is this a fix for a crash or is this just based on codereview?